Embedded Files
The Global Market
The Global Market
The global assistive technology market was valued at USD 21.8 Billion in 2021 and expected to exceed USD 28.8 Billion by 2028 (Vantage Market Research, 2022). Much of the driving force for the growth is due to the increase of disabilities worldwide, in part due to the aging population. According to the World Health Organization (WHO), more than 1 billion people need one or more assistive products globally (WHO, 2018). As a result, the Research & Development for assistive technology presents a growing prospective industry to provide customers with a better quality of life. Since only 1 in 10 in need have access to assistive products (WHO, 2018), there is a demand for industries and governments to fund and prioritize access.
As assistive technologies is a very broad category, providing any data on how AT will grow in the future becomes blurred among the variety of assistive technologies that exist. However, when we start segmenting them by purpose ( into speech-to-text, translation and language learning, etc.) or by customer (learning disabled, language learners, customer service bots, etc), we can see the benefits and trends of certain assistive technologies as they grow in the future.
To narrow down the market, the global market for assistive technology for students with learning disabilities is set to reach $298.13 Million by 2030, an almost doubling from the $154.67million generated in 2020 (Allied Market Research, 2022). This increase can be attributed to the rise in dyslexia numbers, which holds one fourth of the global assistive technology market for students with learning disabilities market. Factors to consider for the surge in prevalence of dyslexia could be due to misdiagnosing. For example, both ADD and dyslexia are auditory processing disorders that share common symptoms and can often be misdiagnosed.
North America continues to dominate the market share by revenue in the global market for assistive technology for learning disabilities as a result of its established healthcare infrastructure and commercialization of products that are supported by government regulations (Allied Market Research, 2022). Though Asia-Pacific is forecasted to have a higher growth in the next period because of its developing healthcare.
Source: PR Newswire
Market Predictions:
Speech-to-text, Text-to-speech and Speech-Enabled Language Translation
Market Predictions:
Speech-to-text, Text-to-speech and Speech-Enabled Language Translation
While it is easy to group these 3 assistive language technologies together (and many to rely on each other to function), their market success and futures are still quite unique.
SPEECH-TO-TEXT
SPEECH-TO-TEXT
Source: Fortune Business Insights
The global speech-to-text market is massive. Measured at $1.3 billion in 2019, it is projected to reach $3.03 billion USD in 2027 assuming an annual rate of growth of 11% during the forecast period
Source: Fortune Business Insights
The goal here is not necessarily for language learning, but for voice enabled applications which leverage machine learning, augmented reality, a natural language processing. Unsurprisingly, this is linked to the growth of the smart mobile device market, smart speaker devices (like Alexa) and the growing tendency to adopt voice-enabled systems. Real-time support services (customer support) and transcription services continue to also be large growth priorities.
Surprisingly, the Covid-19 pandemic has had the opposite effect of this market than it has in other markets. Large enterprises as well as small-to-medium enterprises reduced research and investments in speech-to-text technology. Contrastingly, the demand for such a solutions is expected to see an increase due to the need for social distancing during the pandemic. Looking to the future, healthcare, e-learning, and media/entertainment Industries (for transcription and subtitling) are expected to increase their adoption of this technology.
Source: Fortune Business Insights
The major market drivers here are going to be smart devices and intelligent voice assistants.
The rise of voice assistants like Alexa, Siri, Cortana, and the Google Assistant has increased. Without a doubt, these smart speakers also provide fascinating possibilities for not only interactivity with other networks technologies through application program interfacing (APIs), but also a very real opportunity for companies to data-mine customer communications with an aim towards data analytics. Other analytics that could theoretically be grabbed from voice assistants could include customer profiles measuring things such as voice stress for social analysis and mental well-being. Of course, one of the key points in working with speech-to-text data is permission to gather/use speech-to-text data. Privacy issues pose a very real legal problem.
In terms of market growth, the two highest industries better expected to adopt speech-to-text technologies are those in the customer management area and in the healthcare industry (Amazon already has a healthcare speech-to-text transcriber). Many organizations will be mining there voice data in order to glean insights into customer satisfaction, operational efficiency, and quality across all communications channels. With the additional potential of transcription of a customers audio interaction with an agent, the transcription can then be turned into metadata which can also be analyzed through Natural Language Processing (NLP).
LIMITING FACTOR: Language and linguistic skill.
Ironically, the limitation of NLP technology in interpreting language is language ability in people. Most organizations do not have the necessary framework or components, task-specific lexicon's or semantic interpreters to be able to fully process all of this data. But, to effectively use NLP, the computer system must be able to interpret information across a variety of languages and evaluate and assimilate the expanded conversation. However, this still only works if the individuals who speak the language know what they're talking about: that is, if the workers do not have the task-specific lexicon, the data will not be interpreted appropriately. Additionally, non-native language speakers with mixed proficiency in the language could also dilute data.
Source: Valital
Natural Language Processing:
What is NLP?
Natural Language Processing:
What is NLP?
Natural language processing is a communications bridge between technology and humans. Through NLP, input provided by human voices is understood by computers. It is also considered a branch of artificial intelligence that understands interprets and manipulates human languages. This is link to machine learning approaches, which also require of proper understanding of input data provided by human voice. The growth of NLP has been staggering. It had a $16.5 billion dollar market size in 2020 and is expected to grow to $127 billion by 2028, averaging almost 30% growth per year.
Again, Covid-19 has played a huge part in this. Initially, due to the pandemic, there was a financial downturn in the sector due to poor revenue growth in companies using NLP technology. But, it is quickly becoming very helpful in organizing large volumes of data. In the healthcare industry alone, the increasing use of the internet and connected devices creates tremendous amounts of patient data that can be organized and attract across multiple channels. This ability to filter and organize this data is essentially an assistive technology: computers sorting through data for significant data (Fortune Business Insights, 2022)
TEXT-TO-SPEECH
TEXT-TO-SPEECH
Marketsandmarkets.com specifically looked at the text-to-speech market as impacted by Covid-19. In their report, they value the text-to-speech market at $2 billion US with an estimate to grow by almost 15% to $5 billion US in 2026. Again, handheld devices like smartphones, the growing dependence of elderly populations on technology, increased government spending on education for those who are “differently abled”, and the rising number of people with learning disabilities/ different learning styles who would benefit from speech-to-text technology. Two items of note are mentioned below.
Primary drivers
1) Handheld devices
This is largely fueled by a desire for automation and convenience. Speech-to-text can also substitute for reading particularly as replacements for user manuals: simply responding to verbal prompts from users about problems.
2) Differently abled and different learning styled students
When speech-to-text is integrated into screen readers, it not only offers people the ability to multitask, but also offers those with visual impairment or reading disabilities to use their stronger skills ( listening) to aid learning. Students with disabilities may be subject to annual Awards of 10 to $20,000 or accessible Technologies and Media Services.
LIMITING FACTOR: Proper pronunciation and prosody in naturally occurring speech.
Pronunciation is always been one of the most prominent factors in language learning. If the word doesn't sound correct, the listener may not know its meaning. Speech-to-text technology has always had issues with word stress and tonal/ emphasis differences. There are a large set of rules ( and exceptions) required for properly synthesizing speech in even within the same languages. Additionally, the prosodic nature of the language (the patterns of rhythm and sound) may sound 'off' and could be deincentivizing for people to interact with. Ultimately, these factors up limiting the use of text to speech in various applications.
Source: Markets and Markets
Speech to Speech Translation (SELT)
Speech to Speech Translation (SELT)
According to ExpertMarketResearch in 2020, "The global speech-to-speech translation Market was approximately $335 million US dollars. The market itself is expected to grow at 9.5% annually and by 2026 should be approximately $576 million by 2026 USD.
Source: Expert Market Research
This is tied to a number of factors:
Rising smartphone and smart ‘wearables’ ownership in both developed and developing nations.
Growing integration of AI in speech-to-speech translation.
The increasing user-friendliness of speech-to-speech translation in various apps
growing interests of younger populations to participate cross-culturally
Translation market (SELT)
When describing the global languages services market ( which would include translation, interpretation and speech recognition), MarketResearch.com projected that it would reach about $77 billion US Dollars by 2024. They note trends such as content localization, technological advancement and market consolidation of major players to be impacting on these trends. Unsurprisingly, they also note that Covid-19 has created a favorable market impact, increasing the demand for virtual language service providers.
Modor Intelligence (2021) also points out that, as of 2020, Mandarin Chinese is also gaining high popularity, suggesting that Mandarin may now be the most commonly spoken global language, and the second preferred online application language after English.
Source: Mordor Intelligence
The terms of the competitive landscape, they find that there are very few major players within the market competing to gain stronger segmentation. One of the major upcoming apps they cite is the Google Translate speech-to-speech app for Android-based systems in coming years. Meta/ Facebook is developing a universal speech translator for their 'Metaverse'. Another surprising player is Amazon. They also offer speech translation services through Amazon Transcribe, Amazon Translate and Amazon Polly functioning largely through their Alexa devices: Which can translate one-way from English into 48 other languages.
Source: Mordor Intelligence
Their fastest growing market at the moment is Europe, owing to voice recognition in multi-factor authentication systems, increase in freelance translators and rapid medical tourism across EU member states. They suggest that the largest area of future growth will be the Asia-pacific region, noting the popularity of tourism in the region and the connection between speech-to-speech translation technology and post-pandemic tourism growth.
Source: Mordor Intelligence
Limitations: The present imperfect one-by-one
Compass Magazine (2015) still refers to speech-to-speech as the 'present imperfect': the average error rate has fallen from about 20% incorrect words in 2010 to about 12% in 2013. It was limited to translating single sentences at a time. At that time, speech-to-speech focuses on fairly literal translations of sentences and had difficulties with language surrounding memes, context or conversational ambiguity. It may translate names literally; Mr Kaneda from Japan, is called 'Mr. Goldfield' In English. Surprisingly, not much has changed as Meta reported in 2022 that even with its new AI-powered Universal Speech Translator project (which they piloted with Hokkien and Mandarin Chinese) still is limited to one full sentence at a time.
Activity:
Activity:
Respond to at least one of the two questions in the Padlet below.
Page updated
Google Sites
Report abuse